For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Python function

batch_padded_tokens_and_mask

batch_padded_tokens_and_mask()

max.pipelines.modeling.dataprocessing.batch_padded_tokens_and_mask(start_pos, tokens)

source

Batches input tokens and computes a batched attention mask.

Parameters:

• start_pos (list[int]) – index into the end of the KV cache for each batch item.
• tokens (list[ndarray[tuple[Any, ...], dtype[int64]]]) – unpadded input tokens for this batch.

Returns:

A tuple of (batched tokens, last token indices per batch item, attention mask).

Return type:

tuple[ndarray[tuple[Any, …], dtype[integer[Any]]], ndarray[tuple[Any, …], dtype[integer[Any]]], ndarray[tuple[Any, …], dtype[float32]]]

Was this page helpful?

Thank you! We'll create more content like this.

Thank you for helping us improve!

😔 What went wrong?

Some code doesn’t work

It includes inaccurate information

It's missing information I need

It was difficult to understand

Other

Submit